Author name cluster

Marco Canini

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

8 papers

2 author rows

EWRL Workshop 2025 Workshop Paper

ACING: Actor-Critic for Instruction Learning in Black-Box LLMs

Salma Kharrat
Fares Fourati
Marco Canini

The effectiveness of Large Language Models (LLMs) in solving tasks depends significantly on the quality of their instructions, which often require substantial human effort to craft. This underscores the need for automated instruction optimization. However, optimizing instructions is particularly challenging when working with black-box LLMs, where model parameters and gradients are inaccessible. We introduce ACING, an actor-critic reinforcement learning framework that formulates instruction optimization as a stateless, continuous-action problem, enabling exploration of infinite instruction spaces using only black-box feedback. ACING automatically discovers prompts that outperform human-written prompts in 76\% of instruction-induction tasks, with gains of up to 33 points and a 10-point median improvement over the best automatic baseline in 33 tasks spanning instruction-induction, summarization, and chain-of-thought reasoning. Extensive ablations highlight its robustness and efficiency. An implementation of ACING is available at \url{https: //github. com/salmakh1/ACING}.

PDF

EWRL Workshop 2025 Workshop Paper

Latent Inference for Effective Multi-Agent Reinforcement Learning under Partial Observability

Salma Kharrat
Fares Fourati
Marco Canini
Mohamed-Slim Alouini
Vaneet Aggarwal

Partial observability remains a core challenge in cooperative multi-agent reinforcement learning (MARL), often causing poor coordination and suboptimal policies. We show that state-of-the-art methods fail even in simple settings under partial observability. To address this, we propose LIMARL, a latent-inference framework that augments centralized training with decentralized execution (CTDE) via structured latent representations. LIMARL integrates (i) a state representation module that learns compact global state embeddings, and (ii) a recurrent inference module that enables agents to recover these embeddings from local histories. We provide theoretical analysis on sufficiency and robustness under partial observability. Empirically, LIMARL outperforms strong baselines in diagnostic tasks and challenging SMAC and SMACv2 scenarios, demonstrating better performance and faster convergence. Our results highlight latent inference as an effective and scalable solution for partially observable MARL. An implementation of LIMALR is available at https: //github. com/salmakh1/LIMARL.

PDF

ECAI Conference 2025 Conference Paper

Quantize Once, Train Fast: Allreduce-Compatible Compression with Provable Guarantees

Jihao Xin
Marco Canini
Peter Richtárik
Samuel Horváth

Distributed training enables large-scale deep learning, but suffers from high communication overhead, especially as models and datasets grow. Gradient compression, particularly quantization, is a promising approach to mitigate this bottleneck. However, existing quantization schemes are often incompatible with Allreduce, the dominant communication primitive in distributed deep learning, and many prior solutions rely on heuristics without theoretical guarantees. We introduce Global-QSGD, an Allreduce-compatible gradient quantization method that leverages global norm scaling to reduce communication overhead while preserving accuracy. Global-QSGD is backed by rigorous theoretical analysis, extending standard unbiased compressor frameworks to establish formal convergence guarantees. Additionally, we develop a performance model to evaluate its impact across different hardware configurations. Extensive experiments on NVLink, PCIe, and large-scale cloud environments show that Global-QSGD accelerates distributed training by up to 3. 51× over baseline quantization methods, making it a practical and efficient solution for large-scale deep learning workloads.

Details

ICLR Conference 2025 Conference Paper

Query-based Knowledge Transfer for Heterogeneous Learning Environments

Norah Alballa
Wenxuan Zhang
Ziquan Liu
Ahmed M. Abdelmoniem
Mohamed Elhoseiny
Marco Canini

Decentralized collaborative learning under data heterogeneity and privacy constraints has rapidly advanced. However, existing solutions like federated learning, ensembles, and transfer learning, often fail to adequately serve the unique needs of clients, especially when local data representation is limited. To address this issue, we propose a novel framework called Query-based Knowledge Transfer (QKT) that enables tailored knowledge acquisition to fulfill specific client needs without direct data exchange. It employs a data-free masking strategy to facilitate the communication-efficient query-focused knowledge transformation while refining task-specific parameters to mitigate knowledge interference and forgetting. Our experiments, conducted on both standard and clinical benchmarks, show that QKT significantly outperforms existing collaborative learning methods by an average of 20.91% points in single-class query settings and an average of 14.32% points in multi-class query scenarios. Further analysis and ablation studies reveal that QKT effectively balances the learning of new and existing knowledge, showing strong potential for its application in decentralized learning.

Details

ECAI Conference 2024 Conference Paper

FilFL: Client Filtering for Optimized Client Participation in Federated Learning

Fares Fourati
Salma Kharrat
Vaneet Aggarwal
Mohamed-Slim Alouini
Marco Canini

Federated learning, an emerging machine learning paradigm, enables clients to collaboratively train a model without exchanging local data. Clients participating in the training process significantly impact the convergence rate, learning efficiency, and model generalization. We propose a novel approach, client filtering, to improve model generalization and optimize client participation and training. The proposed method periodically filters available clients to identify a subset that maximizes a combinatorial objective function with an efficient greedy filtering algorithm. Thus, the clients are assessed as a combination rather than individually. We theoretically analyze the convergence of federated learning with client filtering in heterogeneous settings and evaluate its performance across diverse vision and language tasks, including realistic scenarios with time-varying client availability. Our empirical results demonstrate several benefits of our approach, including improved learning efficiency, faster convergence, and up to 10% higher test accuracy than training without client filtering.

Details

ICLR Conference 2021 Conference Paper

AutoLRS: Automatic Learning-Rate Schedule by Bayesian Optimization on the Fly

Yuchen Jin
Tianyi Zhou 0001
Liangyu Zhao
Yibo Zhu 0001
Chuanxiong Guo
Marco Canini
Arvind Krishnamurthy

The learning rate (LR) schedule is one of the most important hyper-parameters needing careful tuning in training DNNs. However, it is also one of the least automated parts of machine learning systems and usually costs significant manual effort and computing. Though there are pre-defined LR schedules and optimizers with adaptive LR, they introduce new hyperparameters that need to be tuned separately for different tasks/datasets. In this paper, we consider the question: Can we automatically tune the LR over the course of training without human involvement? We propose an efficient method, AutoLRS, which automatically optimizes the LR for each training stage by modeling training dynamics. AutoLRS aims to find an LR that minimizes the validation loss, every $\tau$ steps. We formulate it as black-box optimization and solve it by Bayesian optimization (BO). However, collecting training instances for BO requires a system to evaluate each LR queried by BO's acquisition function for $\tau$ steps, which is prohibitively expensive in practice. Instead, we apply each candidate LR for only $\tau'\ll\tau$ steps and train an exponential model to predict the validation loss after $\tau$ steps. This mutual-training process between BO and the exponential model allows us to bound the number of training steps invested in the BO search. We demonstrate the advantages and the generality of AutoLRS through extensive experiments of training DNNs from diverse domains and using different optimizers. The LR schedules auto-generated by AutoLRS leads to a speedup of $1.22\times$, $1.43\times$, and $1.5\times$ when training ResNet-50, Transformer, and BERT, respectively, compared to the LR schedules in their original papers, and an average speedup of $1.31\times$ over state-of-the-art highly tuned LR schedules.

Details

NeurIPS Conference 2021 Conference Paper

Rethinking gradient sparsification as total error minimization

Atal Sahu
Aritra Dutta
Ahmed M. Abdelmoniem
Trambak Banerjee
Marco Canini
Panos Kalnis

Gradient compression is a widely-established remedy to tackle the communication bottleneck in distributed training of large deep neural networks (DNNs). Under the error-feedback framework, Top-$k$ sparsification, sometimes with $k$ as little as 0. 1% of the gradient size, enables training to the same model quality as the uncompressed case for a similar iteration count. From the optimization perspective, we find that Top-$k$ is the communication-optimal sparsifier given a per-iteration $k$ element budget. We argue that to further the benefits of gradient sparsification, especially for DNNs, a different perspective is necessary — one that moves from per-iteration optimality to consider optimality for the entire training. We identify that the total error — the sum of the compression errors for all iterations — encapsulates sparsification throughout training. Then, we propose a communication complexity model that minimizes the total error under a communication budget for the entire training. We find that the hard-threshold sparsifier, a variant of the Top-$k$ sparsifier with $k$ determined by a constant hard-threshold, is the optimal sparsifier for this model. Motivated by this, we provide convex and non-convex convergence analyses for the hard-threshold sparsifier with error-feedback. We show that hard-threshold has the same asymptotic convergence and linear speedup property as SGD in both the case, and unlike with Top-$k$ sparsifier, has no impact due to data-heterogeneity. Our diverse experiments on various DNNs and a logistic regression model demonstrate that the hard-threshold sparsifier is more communication-efficient than Top-$k$.

PDF Details

AAAI Conference 2020 Conference Paper

On the Discrepancy between the Theoretical Analysis and Practical Implementations of Compressed Communication for Distributed Deep Learning

Aritra Dutta
El Houcine Bergou
Ahmed M. Abdelmoniem
Chen-Yu Ho
Atal Narayan Sahu
Marco Canini
Panos Kalnis

Compressed communication, in the form of sparsiﬁcation or quantization of stochastic gradients, is employed to reduce communication costs in distributed data-parallel training of deep neural networks. However, there exists a discrepancy between theory and practice: while theoretical analysis of most existing compression methods assumes compression is applied to the gradients of the entire model, many practical implementations operate individually on the gradients of each layer of the model. In this paper, we prove that layer-wise compression is, in theory, better, because the convergence rate is upper bounded by that of entire-model compression for a wide range of biased and unbiased compression methods. However, despite the theoretical bound, our experimental study of six well-known methods shows that convergence, in practice, may or may not be better, depending on the actual trained model and compression ratio. Our ﬁndings suggest that it would be advantageous for deep learning frameworks to include support for both layerwise and entire-model compression.

PDF Details